While Pods are the fundamental execution units in Kubernetes, managing them individually is not a practical strategy for running applications. A standalone Pod will not be rescheduled if its node fails, and scaling requires you to manually create or delete Pods one by one. This manual process is inefficient and does not provide the resilience expected from a modern container orchestrator.

To automate workload management, Kubernetes uses controllers. A controller is an active reconciliation process that tracks at least one Kubernetes resource type. These objects have a spec field that represents the desired state. The controller's job is to make the current state match the desired state.

This chapter introduces the controllers that manage stateless application workloads. We will cover how to:

• Use a ReplicaSet to ensure a specified number of Pod replicas are always running.
• Manage application rollouts and updates declaratively using Deployments.
• Perform rolling updates to deploy new application versions without downtime and roll back to previous versions when needed.

By the end of this chapter, you will move from managing individual Pods to managing the entire application lifecycle with higher-level abstractions that provide scaling, self-healing, and automated updates.